Microbial cultures:
Cultures of Vibrio harveyi were grown in Tibbles-Rawlings minimal media and cultures of Ruegeria pomeroyi DSS-3 were grown in K media. Agar plates of each axenic culture were inoculated with a freezer stock and visually inspected for culture purity. A single colony from each plate was used to inoculate 50 milliliters (mL) of liquid media in Erlenmeyer flasks, which were incubated in the dark on an orbital shaker at 100 rotations per minute (rpm). Cells were transferred fresh media during mid exponential growth phase (as determined by optical density at 600 nanometers (nm)) for a minimum of two generations prior to incubation for analysis. For O2 utilization experiments, several mL of mid exponential cell culture was aseptically added to several 500 mL serum vials, which were filled with fresh media (filter sterilized to avoid the production of O2-consuming byproducts, equilibrated overnight at room temperature on a stir plate to ensure oxygen isotopic equilibrium with lab air) then sealed with no headspace. The incubations were conducted at room temperatures, which range from 21 to 23 degrees Celsius. A typical incubation included between 4 and 6 bottles, which were sacrificially sampled at each time point. At the desired sampling time, sample liquid containing dissolved O2 was transferred to mercury-poisoned bottle.

DOC experiments:
Dissolved organic carbon (DOC) used in this study was sourced from the Suwannee River, Georgia and Martha's Vineyard Sound, MA, USA. Suwannee River DOC was prepared from a freeze-dried reverse osmosis isolate provided by the International Humic Substance Society (#2R101N; 84% recovery). The isolate was reconstituted in ultra-pure water in a 4-liter (L) precombusted amber glass jug at a concentration of 20 milligrams per liter (mg L-1) (9.7 ± 0.2 mg-C L-1; n = 3). The solutions were adjusted to pH 7.0 ± 0.1 and allowed to equilibrate with the atmosphere on a stir plate for 24 hours prior to filtration with a 0.2 micrometer (μm) Sterivex filter (Milli- pore) and use in photochemical incubations.

Martha's Vineyard Sound seawater was pumped from ∼300 meters (m) offshore (41.530668, -70.645629) at a depth of ∼4 meters into the Environmental Systems Laboratory (ESL) at Woods Hole Oceanographic Institution (WHOI, Woods Hole, MA, USA). At high tide in June of 2021, ∼210 L of seawater was filtered through precleaned (200 L of RO and 100 L of DI water) 0.2 μm ultra-pleat in-line filters (Big Brand Water Filter, Inc.). The seawater DOC was collected in 12 five-gallon, acid-rinsed (10% trace metal grade hydrochloric acid; Fisher Scientific), and precleaned (10x rinses with RO and MQ water) polypropylene buckets. The DOC was isolated using six 5 g PPL cartridges (Agilent Technologies). The methanol eluent from the six cartridges was pooled into one sample. DOC recovery was measured by spiking 100 microliters (μL) of the pooled eluent into a precombusted TOC vial, drying over high-purity N2 (Airgas, Inc.), reconstituting in DI water, and quantifying using a Shimadzu 5000A TOC analyzer. DOC recovery was 44% ± 3% (n = 3). This recovery is ∼15% lower than reported for Brazilian coastal waters. This discrepancy was expected provided the 10× higher C-loading rate used in this study compared to Dittmar et al. (0.4 vs. 0.04 mmol-C g−1 PPL resin), and the well-documented decline in DOC recovery with increased C-loading rates. Martha's Vineyard Sound DOC SPE eluent was subsequently used to create concentrated solutions of coastal DOC (14.6 ± 0.2 mg-C L−1; n = 3). The eluent was added to a 4 L pre-combusted amber glass jug, dried over high-purity N2, and reconstituted in 0.2 μm filtered Martha’s Vineyard Sound seawater. The coastal DOC solutions were allowed to equilibrate with the atmosphere on a stir plate for 24 hours prior to refiltration with a 0.2 μm Sterivex filter and use in photochemical incubations.

The working solutions of terrestrial and coastal DOC were chemically characterized using optical spectroscopy. Specifically, optical proxies for molecular weight and aromaticity were measured [E2:E3; slope ratio (SR); specific UV absorbance at 254 nm (SUVA254)]. UV-visible light absorbance was determined using a Perkin Elmer Lambda 650 s spectrophotometer, whereas DOC concentration was determined as previously described. The optical proxy values determined in this study for terrestrial and coastal marine DOC strongly aligned with those reported in previous studies. The DOC sources were chemically distinct; terrestrial DOC exhibited notably higher molecular weight than coastal DOC.

TOI analysis:
After each incubation (either microbial or photochemical), the seal on each closed bottle was opened and an O2 electrode was inserted for 15 seconds to get an approximate dissolved O2 concentration for later comparison with O2: Ar ratio to ensure no atmospheric contamination. Roughly half of sample liquid was then siphoned out from the bottom of the sample bottle into a preevacuated, prepoisoned (with 500 uL saturated HgCl2 solution) 1 L custom-built glass bottle affixed with a high-vacuum stopcock, similar to previous studies. Method blanks were conducted with O2-free water (bubbled vigorously with helium for 1 hour) to ensure the above methods did not introduce atmospheric contamination. We found the method blank to be sufficiently low as to not influence oxygen isotope measurements within the precision reported in this study (n = 2). The sample bottle was vigorously shaken and allowed to equilibrate overnight. Following equilibration, the sample bottle was inverted, and the degassed sample liquid was evacuated, save 1 to 2 mL of residual liquid to ensure no sample gas was removed. The sample bottle was submerged in a slurry of dry ice and ethanol to freeze all remaining liquid, and the sample gas was introduced to a vacuum line for purification.

Oxygen gas was separated from N2 and Ar using a 3 m gas chromatography (GC) column packed with 5 Å molecular sieve held at -80 degrees C. The integrated peak area (measured by thermal conductivity detector) of the separated O2 and Ar peaks was used to determine the fraction of the starting O2 that remained at each time point. The effluent O2 was collected on silica gel held at at -196 degrees C and passed to one final silica gel cryofocus. The sample was then transferred to a Thermo Scientific MAT 253 Plus isotope ratio mass spectrometer (IRMS).

Each reported analysis is the average value of four acquisition blocks, each consisting of 20 cycles between the reference gas and sample gas (total counting time on sample gas was 400 seconds per acquisition). Measurements were typically analyzed at 5000 mV signal on the m/z 32 Faraday cup (3 × 10^8 resistor). Total acquisition time for a single analysis is roughly 2 hours. All values of δ17O and δ18O were converted to a VSMOW2-SLAP2 scale using a two point calibration of O2 liberated form VSMOW2 and SLAP2 standards using a CoF3 reactor. All slopes (i.e. λ values) and slope uncertainties were calculated as linear fits to plots of δ17O vs. δ18O using the polyfit function in MATLAB (linear least-squares). Typical reproducibility of Ar-free lab air is 0.005‰ for D17O and 0.02 ‰ for δ18O(1σ, n = 10).